The Drosophila gene encoding JIG protein (CG14850) is critical for CrebA nuclear trafficking during development

Abstract Coordination of mitochondrial and nuclear processes is key to the cellular health; however, very little is known about the molecular mechanisms regulating nuclear-mitochondrial crosstalk. Here, we report a novel molecular mechanism controlling the shuttling of CREB (cAMP response element-binding protein) protein complex between mitochondria and nucleoplasm. We show that a previously unknown protein, herein termed as Jig, functions as a tissue-specific and developmental timing-specific coregulator in the CREB pathway. Our results demonstrate that Jig shuttles between mitochondria and nucleoplasm, interacts with CrebA protein and controls its delivery to the nucleus, thus triggering CREB-dependent transcription in nuclear chromatin and mitochondria. Ablating the expression of Jig prevents CrebA from localizing to the nucleoplasm, affecting mitochondrial functioning and morphology and leads to Drosophila developmental arrest at the early third instar larval stage. Together, these results implicate Jig as an essential mediator of nuclear and mitochondrial processes. We also found that Jig belongs to a family of nine similar proteins, each of which has its own tissue- and time-specific expression profile. Thus, our results are the first to describe the molecular mechanism regulating nuclear and mitochondrial processes in a tissue- and time-specific manner.


INTRODUCTION
Among roughly 1500 proteins found in animal mitochondria, onl y 13 pol ypeptides are encoded in the mitochondrial genome of animal cells, including human and fruit fly ( 1 ). The rest are encoded by the nuclear genome, translated in the cytosol and deli v ered to the mitochondria. As a result, many functional protein complexes localized to mitochondria are composed of protein subunits that are encoded by two different genomes. ATP synthase is a typical example of such oligomeric proteins: out of 27 of its subunits, two are encoded in mitochondria, and the rest --by the nuclear genome ( 2 ). Thus, the production of proteins encoded by the nucleus and the mitochondria should be coordina ted. Without proper coordina tion between nuclear and mitochondrial gene expression and the molecular mechanism that establishes mitochondrial-nuclear crosstalk, no ATP synthase and respiratory complexes can be properly generated, leading to a buildup of non-working subunits that would cripple cellular ener getics. The breakdo wn of such crosstalk between two organelles leads to mitochondrial dysfunction, which results in se v ere health issues, such as neurodegenerati v e diseases and cancer ( 3 , 4 ). mosome squash, followed by confocal microscopy experiments, confirmed functional and physical association of Jig with Drosophila CREB (CrebA) in the nuclear chromatin.

Construction of transgenic drosophila
To construct UAS::Jig-GFP, we generated the full-length genomic fragment of Jig locus (Figure 1 A) using PCR. We used wild-type Drosophila genomic DNA as a template for PCR. The r esulting PCR products wer e cloned through The Drosophila Gateway ™ Vector Cloning System (Carnegie Institution of Washington) into the corresponding vector for Drosophila tr ansformation. Gener ating tr ansgenic Drosophila was performed as described ( 19 , 20 ).

Whole mount drosophila tissue immunohistochemistry
Third instar larvae of the appropriate stages were collected prior to dissection. Tissues were dissected in Grace's insect medium, fixed in 4% paraformaldehyde + 0.1% Triton X-100 in PBS for 20 min, and blocked with 0.1% Triton X-100 + 1% BSA for 2 h. These tissues were then incubated  ( 35 ) and Phyre square software ( 36 ). ( C ) Jig protein expressed from embryo to adult stages. Data obtained from the mod ENCODE Temporal Expression Data Project ( 37 ). ( D ) Jig protein expressed almost e xclusi v ely in larval imaginal discs and larval salivary glands. Data obtained from the mod ENCODE Tissue Expression Data Project ( 37 ). ( E ) Two-hybrid approach identified nine proteins interacting with Jig protein in Drosophila ( 38 , 39 ). ( F ) Evolutionary tree of Jig paralogs in the D. melanogaster genome (see also Supplemental Figure S2). Red frames indicate paralogues of Jig located in the same genomic locus. ( G ) Knockdown-transgenes eliminate Jig-GFP protein expression in Dr osophila . Two dif ferent siRNA constructs against Jig were expressed using 69B-GAL4 dri v er in Jig-GFP-e xpressing Drosophila . Total protein extracts from third instar larvae were subjected to Western blot analysis using anti-GFP antibody. Le xA siRNA-e xpressing animals of the same genetic background were used as a control. Tubulin antibody was used as a loading control. ( H ) Jig is r equir ed for Drosophila de v elopment. siRNA against Jig was expressed using 69B-GAL4 dri v er in wild-type Drosophila . All Jig siRNA-expressing animals were arrested in early third instar larval stage. Le xA siRNA-e xpr essing animals of the same genetic background wer e used as a control.

Colocalization analysis
Colocalization analysis was performed on ImageJ ( 22 ) using the Colocalization finder plugin (available at this link: http://questpharma.u-strasbg.fr/html/colocalizationfinder.html ). A 512 × 512 scatterplot was generated based on the CrebA or Jig fluorescence intensity, along chromosome piece. The Pearson correla tion coef ficient was calculated based on this scatterplot.

MitoT r acker r ed staining
Salivary glands from early third instar larvae were obtained through dissection at room temperature. in Phospha te Buf fer Saline (PBS). After a short PBST (0.1%) wash, samples were incubated for 5 min in 100 nM MitoTracker RedCMXRos. After three short washes with PBST (0.1%), the samples were fixed in 4% paraformaldehyde in PBST at room temperature. for 1 min ( 23 ).

Sample pr epar ation f or ChIP-seq
Flies were bred in a tube for an 8-h period, and the eggs laid were allowed to grow at room temperature. The third instar larvae at 12 h stage were collected using 15% sucrose solution. About 0.20 g of larvae was collected. Larvae were washed with 1ml of 1 × PBS by spinning them down at 10 000g. Larvae were homogenized with pestle in 800 ul 1 × PBS, 10ul Protease inhibitor cocktail, 1 ul Tween-20 and 250 ul of PMSF. Formaldehyde was added to 1.8%. Sample was crosslinked on a rotator at r.t.p. for 15 min. Crosslinking was quenched by adding 500 mM of Glycine. The quenched sample was incubated on ice for 5 min. Larval cells were then centrifuged at 1000g for 3 min, the supernatant discarded, and the pellet suspended in 1 ml of Sonica tion buf fer (0.5% SDS, 20 mM Tris, pH 8.0, 2 mM 0.5 M EDTA, 0.5 mM EGTA, 0.5 mM PMSF and 100 × Protease inhibitor cocktail). Samples were then sonicated using the Bioruptor sonication machine for 20 cycles. Sonicated samples were centrifuged at 10 000g for 10 min, and supernatant was collected.
After overnight decrosslinking, 750 ul phenol / chloroform / isoamyl were added to the samples and wer e vortex ed and centrifuged at 10 000g. The top layer was collected, and DNA was precipitated with 1ml of 100% ethanol. Samples were centrifuged at maximum speed for 20 min. Washing was done using 70% ethanol, and centrifuging was performed at maximum speed for 5 min. The pellet was suspended in 22 ul of nuclease-free water.
For each IP, 10% of the sample was used for input. Each IP was diluted to a volume of 1 ml using IP buffer (0.5% Triton X-100, 2mM of EDTA, 20 nM of Tris-HCl pH 8, 150 nM NaCl and 10% glycerol). 100 ul Agarose A beads (50% slurry with IP buffer) were added to each IP. IPs were then rotated at 4 • C for 1 to 2 h. They were centrifuged for 1 min at 1000g. 250 ul of IP buffer and 5 ul of anti-GFP antibody were added to the supernatant. The IPs were rotated overnight at 4 • C. 200 ul of pr otein-A agar ose (50% slurry) were added, and IPs were centrifuged at 1000g in 4 • C for 1min. Pelleted beads were washed with 1ml of low salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, 150 mM NaCl), rotated for 4 min at r.t.p., and centrifuged for 1 minute at 1000g. This washing procedure was repeated 3 times with high salt buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, 500 mM NaCl), 1 time with LiCl buffer (2 mM EDTA, 20 mM Tris-HCl pH 8.0, 0.25 M LiCl and 1% NP-40), and 2 times with 1 ml TE buffer (10 mM Tris-HCl pH8.0 and 1 mM EDTA). DNA from the IPs and Inputs were eluted using 250 ul of elution buffer (1% SDS, 100 mM NaHCO 3 ).
Decrosslinking was done overnight at 65 • C. 15 ul of 1 M Tris-HCl (pH 7.5), 2 ul Gly cob lue and 2 ul Proteinase K were added to each sample and incubated at 65 • C for 30 min. DNA was extracted using 750u l of phenol / chloroform / isoamyl. DNA was sent to Novogene for library preparation and sequencing.

ChIP-seq data analysis
ChIP-seq analysis was done using the w e b-based Galaxy platform. P air ed-end r eads wer e mapped against the DM6 Drosophila melanogaster genomic database. Peak calls were done using the MACS2 callpeak tool with default parameters in Galaxy. Distribution of Jig binding sites relati v e to TSS was generated using the plotheatmap tool in Galaxy with a parameter range set to -3 and +3 kb from TSS ( 24 ).
Gene Ontolo gy anal ysis was done with the String application using the Jig binding gene list ( 25 ). List of acti v e and inacti v e genes during the L3 12 h stage was obtained from Flybase ( 14 ). These gene lists were compared with the list of Jig-bound genes using Excel to determine percent of Jigbound genes that were acti v e genes and inacti v e.
Human orthologs of Jig bound genes were obtained using DIOPT Ortholog Finder from Harvard Medical School ( 26 ). Human CREB target lists were obtained from the CREB transcription factor datasets collected by Harmonizome from Encode Transcription Factor Targets Database ( 27 , 28 ).
Raw data for Drosophila CrebA chip-seq were obtained from the Encode Database ( 28 ). These publicly available deposited data were generated from ChIP-seq on 3 rd instar Drosophila melanogaster larvae with anti-GFP antibody against Cr ebA-eGFP. Cr ebA binding site distribution, target profile and gene ontology data were generated using the same method as that used for Jig.

Electr on micr oscopy
For ultr astructur al analysis, the salivary glands were dissected, fixed in 2% formaldehyde / 2% glutaraldehyde in 0.1 M cacod yla te buf fer pH 7.2, post-fixed in 1% OsO 4 , dehydrated in ethanol and propylenoxide, and embedded in EMbed-812 (EMS, Fort Washington, PA) in flat molds. After polymerization for 60 h at 65 • C, 70nm sections were cut on a Leica Ultracut E microtome (Leica, Austria), placed on collodion / carbon-coated grids, and stained with 2% uranyl aceta te / lead citra te. Sections were viewed on a Tecnai 12 transmission electron microscope (TEM) (FEI, Hillsboro, OR). For EM immunocytochemistry, samples wer e pr epar ed according to Tokuyasu (1980) ( 30 ). In brief, the dissected salivary glands wer e fix ed in 4% formaldeh yde / 0.2% glutaraldeh yde in 0.1M PHEM (60 mM PIPES, 25 mM HEPES, 2 mM MgCl 2 , 10 mM EGTA, pH 6.9), cryo-protected in 2.3 M sucrose, mounted on aluminum pins, and frozen in liquid nitrogen. Thin frozen sections were then cut on a Leica EM UC6 / FC6 cryo-microtome (Leica, Austria), collected on a drop of sucrose / methylcellulose mixture and placed on a formvar-carbon grid. The sections were labeled with primary antibody, and the label was subsequently visualized by colloidal gold conjugated to Protein A. Sections were stained / embedded in 2% methylcellulose / 0.2% uranyl acetate and observed under a Tecnai 12 TEM.

Co-immunoprecipitation assay
Lysa tes for immunoprecipita tion wer e pr epar ed as follows: 15 third-instar larvae were collected for each sample. They were put into Eppendorf tubes and rinsed 3 times with 1 ml of dist. water. 500 ul of ice-cold lysis buffer (10 mM Tris-HCl pH7.5, 150 mM NaCl, 1 mM EDTA, 0.2% NP40, 1% Triton X100, 0.1% SDS, 1% sodium deoxycholate, Complete ™ protease inhibitors (Roche) and 0.1 mM Pefabloc SC (Fluka) were added to each tube, and larvae were homogenized by hand pestle homogenizer on ice. After 30min incubation on ice, samples were centrifuged at 14 500 RPM for 20 min (4 • C). Supernatants were transferred to new Eppendorf tubes on ice. For one immunoprecipitation reaction, 500 mkl of total lysates were incubated with 25 ml Protein-G Sepharose 4B (Sigma #P3296-5ML) on a rotating platf orm f or 1 h 30min at 4 • C. Beads wer e r emoved by spinning 1 min at 2000g. An appropriate amount of antibody was added to the lysates, and the mixture was incubated 4hrs on a rotating platform at 4 • C. The following antibodies were used for immunoprecipitation: anti-GFP (JL8). Then, 30 ul of Pr otein-G Sephar ose 4B were added to the lysates and incuba ted overnight a t 4 • C with rota tion. Beads were washed 4 times for 5 min in 1.0 ml of the lysis buffer. Bound proteins were eluted by 60 ml of 1 × Laemli with heating at 95 • C for 5min.

Third instar larval staging
The start of L3 12 h is defined as the period right after molting into third instar from second instar. This period is ended by a mid-le v el ecdysone spike during the middle stage of Drosophila third instar larval development. This marks the start of puffing stage 1-2 at which time the third instar larvae start wandering out of the food. This time, the larvae have a dark blue gut. This is the mid-stage of third instar larval de v elopment. The latest portion of third instar larval de v elopment is puf f stage 7-9 where dif ferent genes in Drosophila salivary glands become puffed. Larvae have clear guts at this point. This period is ended by a major ecdysone peak causing the larvae to form pre-pupae ( 31 , 32 ).

RN A e xtr action f ollow ed by r everse tr anscriptase qPCR
This assay was performed in triplicate. Twenty third instar larvae were collected for three groups (siRNA Control, siRN A1 JIG and siRN A2 JIG). Total RN A was extracted from cells using the QIAshredder column and RNeasy kit (Qiagen). Contaminating genomic DNA was removed by the g-column provided in the kit. cDNA was obtained by re v erse transcription using M-NLV re v erse transcriptase (In vitrogen). Real-time PCR assa ys were run using SYBR Green master mix (Bio-Rad) and an Applied Biosystems StepOnePlusTM instrument. The amount of DNA was normalized using the difference in threshold cycle (CT) values ( CT) between rpL32 and Jig targets.
Sequences for Jig targets were

els in thir d instar larvae tissues
To e valuated ROS le v els we used MitoSOX ™ Mitochondrial Superoxide Indicators, for li v e-cell imaging (Catalog number: M36006).

NAD and NADH levels in third instar larvae tissues
We used the ScienCell Research Laboratories N AD / N ADH assay (Cat. # 8368).

Motif analysis
Find Individual Motif Occurrences (FIMO) ( 33 ) was used to determine the occurrence of CrebA consensus motif ( 34 ) at Jig binding sites within promoters ( ±1000 bp TSS) using a P -value threshold < 0.001.

Confocal imaging and quantification
Pr epar ed slides were mounted on the Leica TCS SP8 confocal microscope stage and viewed under 63 × optical lens. Using lasers, samples were excited at 488, 552 and 638 nm to detect fluorescence from GFP, CrebA and TOTO3 DNA stain respecti v el y. The w hole cell area was chosen, and the fluor escence was r ecorded using ImageJ. The nuclear area from samples were chosen and fluorescence was recorded. Fluorescence from the nucleus was subtracted from fluorescence recorded from the whole area to get the fluorescence coming from the cytoplasm.
To calculate the distribution of JIG and CREB proteins between the nucleus and cytoplasm, we stained dissected salivary glands with DNA marker TOTO3 and using the appropriate antibodies. Confocal images of the whole organs were taken on Leica DMI8 confocal system and then analyzed using QuPath 0.4.0 software ( 54 ). We utilized the deep-learning neural network StarDist trained to detect fluorescently labeled nuclei ( 55 , 56 ). Nuclei were detected in TOTO3 channel (633 laser), and fluorescence intensity was calculated for proteins (JIG and CREB) in 488 laser channel. We defined the cells as area 10mkm around each detected nucleus. The ratio in protein signal between the cytoplasm and nucleus was be calculated for the whole organ.

Jig encodes a 19kDA protein that controls drosophila development
The Drosophila Jig protein is encoded by a small intron less gene located in the third chromosome (Figure 1 A, Supplemental Figure S2). Jig protein contains 158aa with a nuclear localization signal located at the C-terminus (Figure 1 A, B, Supplemental Figure S1). The Swiss Model ( 35 ) and Phyr e squar e softwar e ( 36 ) pr edicted a 3D model of Jig protein with local protein folding similar to that of the ADPr-cyclase protein (Figure 1 B). The modEN-CODE Tissue and Temporal Expression Project ( 37 ) data analysis demonstrates that Jig expression is limited to a very short developmental stage, 3rd instar larvae (Figure 1 C), and that it is almost e xclusi v ely limited to precursors of adult tissues, imaginal disks, and larval salivary glands (Figure 1 D). Previous genome-wide studies ( 38 , 39 ) reported that CG14850 protein product (Jig) potentially interact with nine Drosophila proteins, including transcriptional factor cyclic-amp response element binding protein A (CrebA), which is involved in nuclear genome regulation ( 40 ) (Figure 1  Owing to its very small size, extensive search does not reveal any obvious homologs to Jig, other than in Drosophila e genomes. In the Drosophila melanogaster genome, Jig has nine paralogs (Figure 1 F, Supplementary Figure S2), four of which are located on the same chromosomal locus (Figure 1 A). Four paralogues, Jig, CG14851, CG8087 and CG13135, share almost all Jig protein features (Supplemental Figure S2A), including fiv e conserv ed cysteines, suggesting that these proteins are either involved in proteinprotein interactions via disulfide bounds or have Zn-fingerlike structural domains.
To study functions and localization of Jig in vivo , we created a transgenic reporter construct by fusing the Jig with a C-terminal green fluorescent protein (GFP) tag under control of inducible UASt promoter (Figure 1 A), and we gener ated tr ansgenic flies expressing the fusion protein (Figure 1 G). The Jig-GFP fusion does not change the phenotype of these flies. They are reproducing, healthy and viable. We also used knockdown transgenic constructs to produce two different nonoverlapping siRNAs. In our control experiment, the expression of these siRNAs effecti v ely eliminated Jig-GFP protein expression (Figure 1 G). The ubiquitous expression of these RNAi transgenes in wild-type Drosophila arrests the fly's de v elopment at early 3rd instar larval stage (Figure 1 H), but does not cause immediate lethality. These observations indica te tha t Jig has a vital function during Drosophila de v elopment.

Jig protein localizes to nuclear chromatin and mitochondria
To monitor the subcellular localization of Jig protein, we expressed the UAS-Jig-GFP transgenic reporter using ubiquitous GAL4 dri v er. An immunob lot analysis using a GFP antibod y ( ␣GFP) demonstra ted tha t the transgene produces a single 46-kDa protein (Figure 1 G). This expression is well tolerated by animals and has no effects on Drosophila de v elopment, viability, fertility or health. Confocal microscopy of dissected tissues of 12hrs third instar larvae expressing recombinant Jig-GFP identified Jig as a protein localized to both nucleus and cytoplasm. In the nucleus, one fraction is bound to chromatin and the other enriched in nucleoli ( We found that intracellular distribution of Jig protein changes during third instar larval development (Figure 3 ). Intrinsic Jig is only expressed from L3 12 h stage to prepupae stage, so we tested only this period. Based on quantification of confocal microscopy pictures, by stage L3 12 h, 73% of the total Jig-GFP had accumulated in mitochondria, while 27% had bound to nuclear chromatin and nucleoli (Figure 3 A). By L3 PfSt 1-2, the distribution had changed to 56% in mitochondria and 44% in nuclei Nucleic Acids Research, 2023, Vol. 51, No. 11 5653 Figure 2. Jig protein localizes to mitochondria and nuclear chromatin in Drosophila third instar larvae cells. ( A ) Jig protein has a dual nuclear-cytoplasmic localization. Jig-GFP (Green) transgene was expressed using 69B-GAL4 dri v er in wild-type 12 h third instar larvae. A single salivary gland cell is shown. DNA was detected using TOTO3 (red) dye staining. ( B ) Jig protein is localized to mitochondria. Third instar larval salivary glands expressing Jig-GFP (green) and mitochondrial protein TIM17B-DsRed ( 15 ) (red) were stained with TOTO3 (Blue) to stain nuclear chromatin. ( C ) Jig protein is localized to mitochondria. Third instar larval salivary glands expr essing Jig-GFP (Gr een) wer e stained with mitotracker568 (red) to detect mitochondria and TOTO3 (red) to stain nuclear chr omatin. ( D ) Jig pr otein binds to nuclear chromatin. Salivary glands were dissected from third instar larvae expressing Jig-GFP, squashed and stained with anti-GFP antibody (Red); DNA was detected using TOTO3 dye (Green). N -nucleus. Scale bars, 15 m. (Figure 3 B). Aside from binding to chromatin and nucleoli, a t this la ter stage in de v elopment, Jig also accumulates in extra-chromosomal bodies (Figure 3 B, arrowheads). By stage L3 PfSt 7-9, 94% of Jig-GFP has alread y transloca ted to nuclei, has been excluded fr om chr omatin and nucleoli, and is now primarily localized to extra-chromosomal bodies (Figure 3 C). Analysis of con-focal microscopy images confirmed the progressi v e relocalization of Jig pr otein fr om mitochondria into nucleoplasm during third instar larval development (Figure 3 D). Taken together, these data strongly suggest that Jig protein may play some role in nuclear-mitochondrial communication and likely coordinates nuclear and mitochondrial functions.

Jig protein is r equir ed f or mitochondria
To test whether Jig plays any role in mitochondrial stability and / or survival, we first analyzed if mitochondrial morphology is affected in absence of Jig using transmission electr on micr oscop y (Figur e 4 A). Tissues from the animals expressing control siRNA and siRNA against Jig were dissected from L3 12hrs stage larvae and compared in respect to mitochondrial morpholo gy. Strikingl y, typical mitochondria (Figure 4 A, left) were scarce in Jig knockdowns (Figure 4 A, right). Instead, we observed a large number of small mitochondria (Figure 4 A, arrowheads), as well as mitochondria with an abnormal phenotype that retained resid-ual cristae inside (Figure 4 A, arrow). Although the morphology of mitochondria is significantly disrupted the functions of mitochondria seems to be not affected (Supplemental Figure S4).
To test whether Jig-knockdown mitochondria are still metabolically acti v e, we employed the JC-1 mitochondrial membrane potential assay ( 29 ), which is used to quantify the fraction of acti v e mitochondria in the cell using confocal microscopy. JC-1 is a fluorescent d ye tha t exists as green-emitting monomers in solution, but these monomers can re v ersib ly aggregate in mitochondria with high membrane potential, forming r ed-emitting complex es and thus

Jig protein is a component of CREB protein complex
CREB protein complex plays important roles in mitochondrial and nuclear transcription ( 11 , 41 ). Two independent groups reported that Drosophila CrebA protein interacted with Jig in yeast two-hybrid ( 38 , 39 ). To confirm the functional interaction of Jig with CREB protein complex in the cell, we first tested if these proteins colocalized in vivo using immunostaining. First, we confirmed that Drosophila CrebA protein localizes to the nuclei and mitochondria (Supplemental Figure S5). To determine if Jig and CrebA colocalize in Dr osophila chroma tin we performed immunostaining of larval polytene chromosomes squash for CrebA and Jig-GFP. Notably, almost 100% of CrebA-positive sites were also occupied by Jig (Figure 5 A) (Supplemental Figures S6 and S7). Co-immunoprecipitation shows that Jig directly interacts with CrebA protein (Figure 5 B). Nuclear localization of CrebA protein, which is normally detected in wild-type Drosophila tissues (Supplemental Figure S5), is se v erely diminished in Jig knockdowns (Figure 5 C, D). In addition, the localization of CerbA protein at the chromatin is se v er ely impair ed in Jig knockdown (Supplemental Figure S8). This data strongly suggests that Jig is r equir ed for CrebA localization at the nucleoplasm.

Jig protein together with CrebA binds promoters in nuclear and mitochondrial genomes
To determine the exact genomic distribution of Jig protein in Drosophila larvae, we performed ChIP-seq assays with anti-GFP antibody. We performed ChIP-seq with wild-type Drosophila as a background control. Analysis of Jig occupancy in the nuclear genome identified two groups of Jig binding sites: unique and repetiti v e genomic sequences (transposons), suggesting that Jig plays a role in the transcriptional regulation of both unique loci and repetiti v e DNA. We identified 1476 Jig binding sites, among which 1469 are in the nucleus and 7 in the mitochondria. We identified 461 sites that Jig bound to be in repetiti v e regions. Similar to CrebA, Jig bounds mostly to the promoter region near the transcriptional start sites (TSSs) (Figure 6 A, Supplemental Figure S9) suggesting that Jig, together with CrebA, is involved in regulation of gene expression. Jig and CrebA binding profile on genomic regions ranging from highly acti v e to silent genes is also the same and they both ( B ) Jig protein interacts with CrebA in Dr osophila . Immunoprecipita tion assays using monoclonal anti-GFP antibod y. Dr osophila stocks expressing Jig-GFP or coexpressing Jig-GFP and Jig siRNA were used. Wild-type (WT) Drosophila stock was used as a control. To detect protein on Western blots, the following antibodies wer e used: rabbit anti-Cr ebA; rabbit anti-GFP (to detect Jig-GFP); rabbit anti-Tubulin. ( C , D ) Jig protein is r equir ed for Cr ebA comple x deli v ery to nuclei. siRNA transgenic constructs against Jig (right panel) or control siRNA (left panel) wer e expr essed using 69B-Gal4 dri v ers in wild-type flies. Salivary glands were dissected from third instar larvae L3 12 h stage and subjected to immunostaining using anti-CrebA antibody (green). DNA was stained using TOTO3 dye (red). The ratio (D) between the green fluorescence intensity in total cytoplasm and total nuclei was calculated for the control and experimental samples. Experiments were performed in three biological replicates. 5-10 cells were analyzed in each experiment. *** P -value ≤ 0.05. Scale bars, 15 m.
bound mostly to acti v e genes (Figure 6 B). We found Jig and CrebA proteins to be broadly bound along the mitochondrial genome and present an identical binding profile (Figure 6 C). Finall y, m utating JIG disrupts the expression of JIG-occupied loci (Supplemental Figure S10). These results suggest that Jig and CrebA bind together to the promoter region of common nuclear loci and to the mitochondrial genome.
Ne xt, we inv estigated the functions of Jig and CrebA target genes. Jig bound to within -5 and +5 kb of 966 genes. Se v enty percent (678) of these genes are acti v e during the first 12 h of third instar larval de v elopment ( 37 ). Interestingly, the Human orthologs of 74.5% of Jig target genes ar e r eported to be CREB targets ( 27 , 28 ). To determine if Jig and CrebA target genes share similar functions we compared gene ontology (GO) of Jig and CrebA target genes. We found that the most enriched functions are common for both Jig and Cr ebA targets (Figur e 6 D). These functions belong to three main categories, de v elopment, morphogenesis and dif ferentia tion, suggesting tha t Jig and Cr ebA r egulate the expression of developmental genes ( 25 ). Finally, se v eral CrebA binding motifs have been identified that all share a similar central sequence of 'CCACGTC' ( 34 , 40 , 42 ). Notably, this motif is present at 69% of Jig-occupied promoters. Collecti v ely, our findings suggest that Jig together with CrebA play a major role in Drosophila de v elopment through gene transcriptional regulation in both nuclear and mitochondrial genome.

DISCUSSION
In this study, we demonstrated that a protein with previously unknown function, which we termed as Jig, localizes to both mitochondria and nucleus, a rare phenomenon demonstrated by only a handful of proteins. As stated before, this type of dual-localized protein is often involved in the communication between mitochondria and nucleus, which is essential for cell survival. Mitochondrion sends signals to the nucleus to transcribe genes that will help to carry out its function and / or to help relie v e stress. Similarly, nu-cleus can also send signals to mitochondria to transcribe genes that will help bring the cell to homeostasis, as needed ( 43 , 44 ). The importance of this nuclear-mitochondrial communication has led many to investigate proteins localized in both mitochondria and nucleus. Howe v er, this communication pathway is still not fully understood. When we identified the Jig protein and determined that it localized to both nuclear and mitochondrial compartments, it prompted a thorough investigation into this protein's potential function in mito-nuclear communication. Through ChIP sequencing and chromosome squash imaging, we showed that this protein not only has dual localization but also binds to the genomes of both nuclear and mitochondrial compartments.
The unique temporal and spatial expression pattern of Jig protein localization is specific to the third larval stage of Drosophila melanogaster de v elopment. Jig mov es from mitochondria to nucleus from early to late third instar larval sta ge. The distrib ution of this protein is highest in the mitochondria at the earlier third larval stage and highest in the nucleus near the end of the third larval stage (Figure 3 ). Jig is vital for Drosophila de v elopment and its knockdown causes de v elopmental arrest in Dr osophila a t third larval stage (Figure 1 H). This de v elopmental stage (larva L3 12 h) at which this arrest occurs coincides with the exclusi v e de v elopmental stage at which Jig is e xpressed in wild type Drosophila (Figure 1 C). Knocking down Jig disrupts normal mitochondrial shapes and sizes, leading to mitochondria with decreased membrane potential (Figure 4 ). We also showed that Jig does not function alone. It binds to Drosophila CREB. This protein has been a subject of immense study in mammals for understanding its role in nuclear and mitochondrial communication. Ever since the discovery of retrograde response genes (RTGs) in yeast that mediate mitochondria signaling to the nucleus, thus establishing a retrograde communication pathway with it, researchers have been trying to find equivalent genes in mammals. Instead, in mammals, they found that this function is served by several genes that establish a retrograde communica tion signaling pa thway between mitochondria and the nucleus. CREB is one of these proteins ( 12 ).
Under stead y sta te condition, CREB transloca tes to mitochondria ( 45 , 46 ). When a lot of changes are happening inside a cell, like during de v elopment, mitochondria send signal to nucleus through translocation of nuclear transcription factors like CREB to the nucleus to regulate genes ( 47 ). Here we showed that Jig protein binds to CREB and colocalizes with CREB in the nuclear chromatin, whereas Jig knockdown disrupts CREB localization into the nucleus (Figure 5 C, D). When mitochondria need to communicate with the nucleus, the Jig protein 'hooks' CREB and facilita tes its traf ficking to the nucleus where CREB can bind to genes to regulate nuclear / mitochondrial functions. Thus, any dysfunction in mitochondria will lead to retrograde signaling carried by proteins like CREB to help return cells to homeostasis. Ther efor e, it is not surprising that knockdown of Jig pre v ents CREB accumulation in the nucleus and causes mitochondria to lose their functionality and become smaller and less acti v e. These compromised mitochondria are energetically impaired by the loss of nuclear communication. This leads to e v entual arrest of growth and lethality in Jig knockdown Drosophila .
Both Jig and CrebA bound to almost the whole mitochondrial genome in exactly the same pattern, suggesting a Jig-CrebA dependent regulation of these genes. These genes code for subunits of Oxidati v e phosphoryla tion pa thway proteins ( 48 ). This explains why knockdown of Jig causes morphological and functional changes in mitochondria (Figure 4 A-C). Interestingly, despite the morphology of mitochondria that is significantly disrupted, the functions of mitochondria do not seem affected (Supplemental Figure S4). This result could be explained by a global downregulation of metabolism during the third instar lar-vae to pr epar e metamorphosis (49)(50)(51), that could mask the effect of Jig on mitochondrial metabolism. Furthermore, we showed that Jig and CrebA bind together mostly to acti v e genes (Figure 6 B). These genes are mainly involved in developmental processes (Figure 6 D). Developmental arrest during third instar larval stage that we observed when Jig function is disrupted could be due CrebA no longer being transported to the nucleus, leading to a misregulation of the expression of these developmental genes (Figures 1 H, 6 D). In parallel, we identified loci where Jig binds, while CrebA is absent. This result could re v eal a second role of Jig in the regulation of gene expression that is CrebA-independent.
Even though it performs a vital function, Jig is only present in Drosophila during a specific de v elopmental stage. We predict that this vital function is carried out by its paralo gues (Figure 1 F) w hen Jig is no longer pr esent. P aralogues, such as CG11300, CG14852 and CG12491, are expressed in either the first larval or the embryonic stage too, potentially carrying out a function similar to that of Jig protein ( 14 ). Jig protein has a conserved region of 5 cysteines that are also present in a group of its paralogues: CG14851, CG13135 and CG8087 (Supplemental Figure S2). Interestingly, these are not the only proteins with these conserved cysteine groups since they are also found in the Arabidopsis plant species. This group of proteins is named as Plasmodesmata callose binding proteins (PDCBs) because they function in cell-cell trafficking by anchoring to the plasma membrane using the part of the structure that contains the 5-cysteine group ( 52 ). One possible interpretation of the structur e / function r ela tionship in the Jig protein is tha t the cysteine-based motif helps in anchoring CREB to DNA. Along with that feature, Jig has a strong positi v ely charged Arginine region at its C-terminal end which helps it to localize to the nucleus. It is well known that CREB forms a complex with other proteins like CREB binding protein, ATF1, in the region of the gene it is going to transcribe ( 53 ). Because of its structure, our model suggests that Jig plays an anchoring role for this complex on the gene promoter.
Jig interacts with other TFs, such as LOLA and XBP1. LOLA is important for neural de v elopment of Drosophila during the embryonic stage ( 54 ). Its knockdown leads to problems with axonal growth causing lethality. XBP1 is involved in the unfolded protein response (UPR) pathway ( 55 ). It transcribes genes that can degrade and / or fold the unfolded proteins in endoplasmic reticulum. Jig also interacts with SKPA, another protein involved in protein degradation. SKPA degrades ubiquitin-tagged proteins ( 56 ). Interestingly, SKPA knockdown also causes lethality owing to motor dysfunction. The disappearance of Jig after third larval stage, together with its interaction with XBP1 and SKPA, suggests its potential degradation via the UPR pathway, or through ubiquitination and degradation by SKPA, after Jig is done deli v ering CREB to the nucleus and helping it to transcribe the necessary genes. Other interactors of Jig, such as ODA and DIDUM, are also vital for Drosophila development and, similar to Jig knockdown, knocking down any of them results in lethality before the end of larval stage ( 31 ). Interestingly, Jig interactors show a bias towar ds neural de v elopment. So, it is possib le that Jig, along with its function to traffic and anchor CREB to nucleus, also functions in neural de v elopment pathways through its